Title: M-String Segmentation: a Refined Abstract Domain for String Analysis in C programs
Location: Meeting room B, Building Zeta
Type: Research Result
Speaker: Martina Olliaro
Abstract: We present a refined segmentation abstract domain for the analysis of strings in the C programming language, properly extending the parametric segmentation approach to array representation introduced by P. Cousot et al. to the case of text values. In particular, we capture the so-called string of interest of an array of char, in order to distinguish well-formed string arrays. A concrete and abstract semantics of the main C header file string.h functions are worked out in full detail.
Title: Lw: a new general-purpose programming language
Location: Meeting room, Building Zeta
Type: Research Result
Speaker: Alvise Spanò
e introduce Lw, a new general purpose, statically typed, strict, impure, functional language supporting cutting-edge features and advanced forms of polymorphism for writing robust, reusable and succinct code. It integrates state-of-the-art advancements in the field of programming languages together with a number of novel bits which makes it ideal for writing big as well as small programs: each heavyweight declarative language construct offers an inferred lightweight counterpart, allowing programmers to design large software architectures that seamlessly coexist with more script-like code.
Among its highlights: type and kind inference, System-F types and first-class polymorphism, open-world overloading with automatic context-dependant resolution, implicit function parameters and controlled dynamic scoping, Generalized Algebraic Datatypes (GADTs), row types for polymorphic variants and records, powerful kind system supporting higher-order polymorphism and kind polymorphism, first-class modules and much more.
Resolution of type constraints is central to many language mechanisms, which, combined with overloading, leads to a form of static dispatching that can either be automatic or assisted by the programmer; dually, row-typed records are subject to dynamic dispatching by nature and enables structural subtyping – a.k.a. sound duck typing. And here lies one of Lw’s most notable and novel features: users can turn type constraints into records and viceversa anytime by using a pair of special inject/eject operators, converting a non-first-class entity which basically resembles a dictionary into a first-class record value, and the other way round. This makes two worlds communicate: the world of static resolution and the world of dynamic resolution. Languages out there typically do not define a clear symmetry in this respect; plus, a lot of boilterplate code is often required for switching between the two worlds, when possible at all.
In Lw this symmetry is crucial and explitly designed, encouraging code reuse.
Title: Building Domain Specific Languages with the Metacasanova meta-compiler
Location: Meeting room
Type: Research Result
Speaker: Francesco di Giacomo
Abstract: main specific languages (DSL’s) are becoming more popular thanks to their ability to provide abstractions that are not part of general purpose languages that ease the development process in specific scenarios, ranging from video games to web development.
Creating a new DSL requires to build a compiler for it, which is a very complex piece of software. This complexity usually causes two problems: (i) the development of the compiler itself requires a significant amount of manpower, and (ii) adding new features to the language may become problematic. In this talk we propose a different approach to writing a hard-coded compiler by using a meta-compiler, i.e. a software that takes as an input the language definition and a program written in that language, and outputs a corresponding executable code. We show the advantages of this approach in terms of simplicity of design and coding requirements, and we present a performance analysis on the generated code for Casanova 2, a DSL for game software development.
Title: The evolutionary ecology of programming languages
Location: Meeting room, building Zeta
Type: Research result
Speaker: Silvia Crafa
Abstract: Taking advantage from the theoretical understanding of biological evolution, I will review the history of popular programming languages from an evolutionary perspective, focusing on the quest for good programming abstractions. In particular, I will discuss the different abstraction levels involved in present-day concurrent and distributed programming, commenting on the role of the theoretical approach. The general aim is pointing out a number of remarks in order to grasp a unifying, but not simplistic, view of programming languages development.